Scribd
Upload
21 views9 pages

PPNR Modeling FRB New York Kovner

PPNR modeling uses historical relationships between revenue components, macroeconomic variables, and asset balances to project revenues. A basic approach estimates these relationships for a panel of banks, inputs a macroeconomic scenario, and forecasts balances to generate PPNR projections. Key aspects include using pro forma or ratio data to account for acquisitions, allowing coefficients to differ across banks, and choosing specifications grounded in historical performance and economic intuition. The model must be tested against actuals from different periods, including recessions, to evaluate its ability to forecast under stress.

Uploaded by

Srinivasa
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
21 views9 pages

PPNR Modeling FRB New York Kovner

PPNR modeling uses historical relationships between revenue components, macroeconomic variables, and asset balances to project revenues. A basic approach estimates these relationships for a panel of banks, inputs a macroeconomic scenario, and forecasts balances to generate PPNR projections. Key aspects include using pro forma or ratio data to account for acquisitions, allowing coefficients to differ across banks, and choosing specifications grounded in historical performance and economic intuition. The model must be tested against actuals from different periods, including recessions, to evaluate its ability to forecast under stress.

Uploaded by

Srinivasa
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 9

PPNR Modeling

Anna Kovner, FRB New York

The views expressed here are my own and not necessarily those of the Federal
Reserve or its staff. 1
An econometric approach to PPNR modeling
• Begin with a historical time series of actual or pro forma data
assuming that all merged entities were part of the ultimate
acquirer as of the beginning of the sample
– Pro forma allows for firm fixed effects and autoregressive model
• Macro variables based on those forecast in scenario
– Can add other variables, but then you need to know how they vary
with those forecast in scenario
• Historical relationship between macro variables and revenue
components (normalized by asset balances) is estimated and
projected into the future for a panel of banks
• Firm specific differences arise in model by: controlling for
bank asset allocation, fixed effects, allowing coefficients to
differ with bank or bank type 2
Predicting PPNR
• Basic idea: [1] estimate historical relationships; [2] input macro scenario; [3]
estimate evolution of balances; [4] plug into each model to generate
forecasts; [5] combine to produce PPNR projection.

Historical data on
PPNR components Forecast macro
scenarios Forecast
PPNR
Assets
PPNR
Historical macro data Estimated coefficients (Dollar
Amount)

Historical data on Asset allocation


asset balances balance forecast Asset balance
forecast

Estimated coefficients Forecast ratio in each


category
3
Model specification overview
• Macro variables
– Different variables matter for different components of
PPNR
What can make this approach firm specific?
• Lagged dependent variable i.e. the dependent variable in
question for the same firm in the previous calendar quarter
• Controls for firm characteristics: size and the composition of
assets
• Can include BHC dummies or fixed effects
• Can separately estimate for different firms or firm types

4
How to capture acquisitions? Pro forma
• Pros
– Accounts for current geographic reach and diversification
of business
– Does not confuse trends generated by acquisitions with
organic business trends
• Cons
– Summation of separate businesses does not allow for
economies of scale or changes to integrated businesses
– Data can be difficult to get

5
How to capture business changes? Ratios
• Pros
– Smooths acquisitions
– Does not confuse trends generated by acquisitions with
organic business trends
– Allows for revenues to be associated with appropriate
asset investment
– Captures changes in business focus
• Cons
– Some revenue streams may not vary with balance sheet
assets (i-banking, asset management)
– Requires link between PPNR and balance sheet models for
internal consistency 6
Model choices – Levels of ratios
• Pros
– AR captures persistence of revenue
– Prevents model from generating unrealistic results
because model is grounded in historical level rather than
change
– Can converge to historical mean (bank level or industry
level)
• Cons
– False sense of security from high r-squared
• Other choices: changes in levels, changes in levels of ratios

7
What’s a good model?
• Based on data
– When judgment replaces data, should be a good reason
• Consistent with historical performance
• Consistent with recent trends in industry and
bank-specific
• Consistent with economic intuition – revenue
should not increase in a crisis, recent bad trends
should not become good in a crisis, etc.
• Good budgeting for normal times may not be a
good framework for predicting stress outcomes

8
What’s a good model?
Test Pros Cons
Adj-R squared • Methodologically sound • Easy to over fit with limited
data
• May produce odd results if
macro variables are
correlated
vs. actual realizations (1Q ahead) • Traditional forecasting test • Good performance in
baseline ≠ good performance
in stress
vs. 2007-8 actual realization • Relevant stressed data • Business may have changed

vs. actual • Relevant stressed data • Business may have changed


in recession

Base vs. Stress • Reality check • Hard to know what the


absolute level of difference
should be

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy